INDEX

Explanations

phrases expressing deep understanding or realization

oai_token-act-pair · gpt-3.5-turbo

Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res_fs24576-jb

Configuration

jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_24576

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

24,576

Data Type

torch.float32

Hook Point

blocks.8.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

IFrame

Link

Negative Logits

vice

-0.70

States

-0.66

 Allied

-0.64

uel

-0.61

Britain

-0.61

 Julius

-0.60

 Proceedings

-0.59

rouse

-0.58

Mos

-0.56

Nim

-0.56

POSITIVE LOGITS

 enough

0.84

rontal

0.77

ãĤ¦ãĤ¹

0.76

ortium

0.74

ãģ®å

0.74

ãĥĨ

0.73

umerable

0.72

 plenty

0.72

 Enough

0.71

Enough

0.71

Activations Density 0.162%

phrases expressing deep understanding or realization

No Known Activations